Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly

Sign up
Appearance settings

fix(init): default scheduler.backend to file in scaffolded forge.yaml#180

Merged
initializ-mk merged 1 commit into
main from
fix/init-scheduler-file-backend
Jun 16, 2026
Merged

fix(init): default scheduler.backend to file in scaffolded forge.yaml #180
initializ-mk merged 1 commit into
main from
fix/init-scheduler-file-backend

Conversation

@naveen-kurra

@naveen-kurra naveen-kurra commented Jun 16, 2026

Copy link
Copy Markdown
Collaborator

Problem

Follow-up to #177 (this change was authored alongside it but didn't land in the merge).

Once an agent has a working LLM executor, the runtime initializes the scheduler. SchedulerConfig.Backend defaults to auto, which selects the Kubernetes CronJob backend when running in-cluster — and that errors at startup unless scheduler.kubernetes.service_url is set:

Error: kubernetes scheduler backend: scheduler.kubernetes.service_url is required

Platform-scaffolded agents deploy via their own k8s manifests (not forge package, which would set service_url), so the auto→kubernetes path crashlooped (CrashLoopBackOff) as soon as a real executor came up.

Fix

Default the init template to the in-process file scheduler:

scheduler:
 backend: file

Opt into durable K8s CronJob scheduling by setting backend: kubernetes + kubernetes.service_url.

Verified

Live deploy (testing-sidecar5): pod reaches 2/2 Ready with no restarts; agent logs using LLM executor (no scheduler crash, no stub). Combined with #177, the full platform path — secret load → LLM executor → scheduler — now works end to end.

🤖 Generated with Claude Code

Without an explicit backend, SchedulerConfig defaults to 'auto', which
selects the Kubernetes CronJob backend when running in-cluster — and
that errors at startup with 'kubernetes scheduler backend:
scheduler.kubernetes.service_url is required' unless service_url is set.
Platform-scaffolded agents are deployed in-cluster via their own k8s
manifests (not 'forge package', which would set service_url), so the
auto/kubernetes path crashlooped once a real LLM executor initialized
the scheduler. Default to the in-process file backend; opt into the
Kubernetes backend by setting backend: kubernetes + kubernetes.service_url.
@initializ-mk initializ-mk merged commit fd929a2 into main Jun 16, 2026
9 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

None yet

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

AltStyle によって変換されたページ (->オリジナル) /